Get the latest tech news
Mastering Atari Games with Natural Intelligence
The blog post explores the groundbreaking achievements of Genius™-powered agents in mastering Atari games, significantly surpassing the performance of leading AI models while using 90% less data. These agents, based on Bayesian Inference and Active Inference principles, demonstrated superior competency in various games with minimal training data and computational resources, emphasizing a shift towards more efficient and generalizable AI solutions. This milestone showcases the potential of neuroscience-based methodologies in advancing machine intelligence.
In order to provide an apples-to-apples comparison against state-of-the-art (SOTA) machine learning, for these initial tests, we selected model-based, IRIS, which is based on the breakthrough transformer architecture which in turn is the foundation of Generative AI and LLMs like GPT, Claude, Gemini, Llama, Grok and others. In the world of statistics, machine learning, and artificial intelligence, Bayesian inference is viewed as a powerful and elegant framework due to its principled, probabilistic methods for reasoning under uncertainty, but the computational demands have made it challenging to implement beyond toy problems until now. Rather than a few incomprehensibly costly yet unreliable models controlled by a handful of corporations, imagine trillions of inexpensive, hyper-efficient, specialized autonomous self-organizing agents operating at the edge and in the cloud, coordinating and collaborating with a common intrinsic – and deceptively simple goal at all levels from individual to the collective, of seeking to understand, i.e. to reduce uncertainty.
Or read this on Hacker News